ResearchTasksFree-form answer generationFree-form answer generationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedKLAR (OOD)GRPO38.13Accuracy12Jun 8, 2026KLAR (train)GRPO49.69Accuracy12Jun 8, 2026